Findings from lab experiments in urban and rural Uganda

Transcription

1 Gender, context and competition: Findings from lab experiments in urban and rural Uganda Kjetil Bjorvatn a *, Ranveig Falch b, Ulrikke Hernæs c abc Department of Economics, NHH Norwegian School of Economics, Helleveien 30, 5045 Bergen, Norway ABSTRACT We analyze the willingness to compete among secondary school students in an urban and a rural context in Uganda, and find no significant gender bias in competitiveness in the urban context, while females in the rural context are far less competitive than males. In contrast, the willingness to compete among males is not sensitive to location. This result holds when controlling for knowledge, confidence, attitude towards risk, and a number of individual and household background variables. Our analysis suggests that context is an important determinant of willingness to compete, and we speculate that attitudes towards gender equality may be an important factor in shaping these preferences for females. JEL classification: D03, O12 Keywords: Competition; Gender; Urban and rural; Development * Corresponding author at NHH Norwegian School of Economics, Bergen, Norway. address: Kjetil.Bjorvatn@nhh.no (K. Bjorvatn). Phone number:

2 1. Introduction Willingness to compete is generally viewed as important for success in the labor market. Indeed, competitiveness measured in the lab has been found to correlate with competitive choices and successful outcomes in the field (see Buser et al., 2014, Zhang, 2012 and 2013, and Berge et al., 2014). Many studies find that females are less competitive than males, and ascribe at least part of the observed gender gaps in labor market outcomes to differences in competitiveness (see Croson and Gneezy et al, 2009; Niederle and Vesterlund, 2007). However, the gender gap in competitiveness does not seem to be universal: it appears to be conditional on context and not biologically fixed. For instance, Gneezy et al. (2009) show that females in a matrilineal community in India are more competitive than males, and Zhang (2012) that there is no gender difference in competitiveness within the majority population, which the author ascribes to the majority s exposure to communist ideology. Booth and Nolen (2012) demonstrate that girls from all girls schools in the UK are as competitive as boys from coeducational schools, and more competitive than girls from co educational schools. In sum, the level of female empowerment appears to be an important determinant for women s willingness to compete. 1 Our paper provides evidence on gender and competitiveness from a series of labexperiments in rural and urban Uganda. To the best of our knowledge, this is the first experimental paper studying gender and competition in an urban and a rural setting. 2 Rural development is high on the policy agenda in many developing countries, and, given the importance assigned to competitiveness for labor market 1 Observe, however, that Cardenas et al. (2011) find that females in urban Colombia are as competitive as males, while females in Sweden are less competitive than their male counterparts. Given that the level of gender equality is very high in Sweden, this is somewhat surprising. 2 Douoguih (2011) studies competitiveness in a rural and urban context in Ghana. Around 30 per cent of the participants in her experiment chose to compete, and more so in the urban than in the rural context. She does not, however, discuss gender differences. 2

3 outcomes, providing evidence on differences in competitiveness in the urban and rural context could have important policy implications. Attitudes towards gender equality are very different in the rural and urban setting, and hence it is a natural place to look for evidence on whether and how the context shapes the willingness to compete. There is ample evidence that female discrimination in Uganda is particularly pronounced in the rural areas. One indication of this is the gender gap in literacy rates, which is much higher in the rural area than in the urban area. According to the Uganda National Household Survey 2009/2010 (UBOS, 2010) 90 percent for males 86 percent for females in urban setting, the corresponding numbers in the rural setting being 77 percent and 62 percent. This survey also reports on views on gender equality, which show a similar pattern. Among the females themselves, 61 percent in the rural area (aged 15 49) agrees that the husband is justified in beating the wife in case she burns the food, argues with him, goes out without telling him, neglects the children, or refuses to have sexual intercourse with him. The corresponding number among urban women is 46 percent. While the acceptance of wife beating is of course shockingly high also in the urban area, the situation is clearly even worse in the rural context. This pattern is also present in financial matters, where in the rural setting only 49 percent of wives control their own earnings, with the corresponding number in the urban setting being 67 percent. It is plausible that these attitudes towards gender equality may also affect women s willingness to compete, as they have been shown to do in the studies on gender and competitiveness from India, Tanzania, China and the UK referred to above. Indeed, individualism has been found to be an important factor in determining competitiveness (Leibbrandt, Gneezy and List, 2013), and male dominance gives little room for female individualism. If so, then a policy advice would be to strengthen the willingness to compete among young women in the rural area in order to improve their livelihoods. 3

4 Figure 1 gives a preliminary glance at our data among secondary school students in Uganda, showing the share of participants, by gender and location, who chose to compete on an incentivized multiple choice test in the lab. It is clear that in the urban location, females compete almost as much as males (35 percent vs. 40 percent), while in the rural context, females compete less than half as much as males (16 percent vs. 37 percent). Of course, this is just raw data, and we need to take into account differences in knowledge, confidence, attitudes towards risk, and other background variables that are likely to explain competitiveness. We return to this issue in section 4. Figure 1. Competition by gender and location Note: The Figure shows willingness to compete, measured as the choice of a competition rate rather than a fixed rate in an incentivized multiple choice test, and the 95% confidence intervals. The rest of the article is organized as follows: Section 2 describes the participants, section 3 introduces the lab experiment, section 4 contains the analysis, and section 5 concludes. 2. The participants The participants consisted of 780 secondary school students in Uganda, of which 381 4

5 were recruited from an urban setting (Kampala) and 399 from a rural setting (Tororo). Kampala is the capital with a population of 1.5 million, while Tororo is a rural district in eastern Uganda with a population of 450,000 of which approximately 90% live in rural sub counties. 3 The main ambition of the lab experiment was to investigate attitudes towards disability among secondary school students, and hence involved around 25 30% students with a physical impairment in each session (see Bjorvatn and Tungodden, 2014). We randomly selected schools to participate in the experiment. 4 The school administration was informed that this was a research project on youth and entrepreneurship and was asked to provide us with a representative sample of students for the experiment. Overall, 32 schools in the urban region and 36 schools in the rural region were recruited to participate in a total of 20 sessions, ten held in each location. Table 1 presents averages of the background variables of interest: Gender, age, whether the participant was disabled, whether they had a TV at home, and how many times per week they had meat for dinner. Column (1) shows the numbers for all the participants, in column (2) and (3) these are differentiated by location, and the difference between the urban and rural setting is shown in column (4). The sample is balanced on gender in the two locations, while the participants in rural setting are, on average, eight months older than those in the urban setting. It is less common to have a TV in the rural setting compared to the urban one, and meat consumption, which is a proxy for wealth, is the same in both locations. 3 Numbers from (accessed ) and the Tororo District Data Base: 4 For practical reasons, in Kampala students with disabilities were recruited from schools in the vicinity of the lab-venue. Neither disability nor priming, which was used to elicit the impact of social identity on attitudes and behavior, predict competition choice, and including these as controls do not affect the qualitative results. 5

6 Table 1. Sample overview (1) (2) (3) (4) All Urban Rural Difference Female (0.500) (0.500) (0.501) (0.036) Age *** (0.482) (0.475) (0.488) (0.034) Disabled ** (0.49) (0.47) (0.50) (0.04) TV *** (0.498) (0.476) (0.438) (0.033) Meat (1.404) (1.507) (1.301) (0.101) Number of observations Note: Columns (1) (3) report average values for different background variables, with standard deviations in parenthesis. Female is an indicator variable that takes the value of 1 if the participant is a female. Age shows the participant s age in number of years. Disabled takes the value of 1 if participant has a disability. TV is an indicator variable taking the value of 1 if the participant has a TV at home. Meat is the number of times the participant eat meat in a week. Column (4) shows the differences between the urban and the rural setting with a t test for equality and standard errors in parenthesis, *p<0.10, **p<0.05, ***p<0.01. There were ten sessions in the urban location (1 10) and ten sessions in the rural location (11 20). Figure 2 shows attendance across sessions. The average number of participants was 39, with a minimum of 33 (session 8), and a maximum of 42 (sessions 3 and 14). The standard deviation is

7 Figure 2. Attendance by session: urban (1 10) and rural (11 20). Note: Figure shows the number of participants in each session (1 20). Figure 3 shows the gender composition by session. As documented in Table 1, in the lab sessions as a whole, an average of 51 percent of the participants were females (52 percent in the urban location and 50 percent in the rural location). While we sought to achieve gender balance in each session, there was some variation, with a minimum of 37 percent females (session 4) and a maximum of 63 percent females (session 7), the standard deviation being

8 Figure 3 Share of females per session: urban sessions, 1 10; rural sessions, Note: The Figure shows the share of female participants in each session (1 20). 3. The lab All tasks were incentivized to ensure that the participants answers reflected their abilities, beliefs and preferences. 5 Willingness to compete was measured by giving the participants a choice between a fixed and a competitive rate on a set of multiplechoice questions on general topics and mathematics; all topics, as well as this form of testing, were familiar to the secondary school students in our experiment. For each question, there were four possible answers, of which one was correct. An example of a general topic question is the following: Which country has the largest population?, with the four possible answers being India, USA, China, or Brazil. An example of mathematical question was: Multiply 34 by 238, with possible answers of 8082, 8032, 8092, or The participants were informed of their earnings and paid in cash in sealed envelopes at the end of their session. The participants were also informed of this procedure at the beginning of the experiment. The instructions from the lab are included in the Appendix. The part of the lab relevant for the present paper is mainly 4.3 Beliefs and competition choice, 4.1 Objective questions: Round 1, and 3.2 Risk. The Appendix also includes screenshots of the Beliefs and competition choice, and for Risk. 8

9 There were two rounds of multiple choice questions. In the first round, all of the participants were offered a fixed rate of 200 Ugandan Shillings (Ush), which corresponds to around 0.1 USD, for each correct answer. Prior to the second round, the participants were asked whether they wanted a fixed rate, similar to that of the previous round, or a competitive rate. The competitive rate would ensure a 500 Ush payment for each correct answer, given that the participant got at least as many correct answers as the entire session average in the first round, and 0 Ush if the participant got less than the average. 6 In the analysis, we measure competition by an indicator variable taking the value of 1 if the participants choose to compete, and 0 if the fixed rate was chosen. The questions in both rounds were similar. In the second round, the participants were rewarded according to their chosen contract. Figure 4. Knowledge, by gender and location. Note: The Figure shows the number of correct answers (1 10) on the first round (fixed rate) of multiple choice questions, and the 95% confidence intervals. Figure 4 shows the performance on the first round of multiple choice questions, by 6 The moderator and lab-assistants made sure that instructions were well understood before the participants made their choices or carried out the tasks. 9

10 gender and location. The results show that the females have weaker performance than the males, and the females in the rural area (5 correct answers) received the lowest score (average score for full sample is 6). From the results in Figure 4 it is obvious that when analyzing willingness to compete, we need to control for the test scores (Knowledge, 0 10). In addition, we control for risk preferences (Risk, 0 3), measured by the number of times the participant chose the uncertain alternative in a gamble between a guaranteed sum of 2000 Ush, and the probability of winning 4000 Ush with 25, 50 and 75 per cent, respectively. 7 Figure 5 shows the number of risky choices by gender and location (0 3). The results reveal that the willingness to take risk is similar across genders but different across contexts; people in the urban location are more willing to make the risky choice (1.2 times out of three) than those in the rural location (one time out of three). Figure 5. Risk, by gender and location. Note: The Figure shows the number of risky choices (0 3), and the 95% confidence intervals. 7 The participants were informed that one of the cases would be randomly chosen to determine their payments. 10

11 We also control for Confidence ( 10,10), measured after the first round of multiplechoice questions. Confidence is a defined as the participants stated beliefs about their own performance in the first round relative to the performance of the other participants in the session. For instance, if a participant guessed that own performance was 5 and guessed that the average performance was 6, the confidence measure would be minus 1. The participants were told they would earn a reward of 1000 Ush for each correct guess. Figure 6 reveals that the participants in general guessed that their own performance was better than the average participant in the session, with the exception of the rural females who believed that their performance was worse than the average (which we know is also true). Urban males were the most confident. Figure 6. Confidence, by gender and location. Note: The Figure shows confidence, measured as the difference between beliefs about own performance and the average performance in the session ( 10,10), and the 95% confidence intervals. 4. Analysis Table 2 shows the results of the OLS regressions, willingness to compete is the 11

12 dependent variable. 8 Columns (1) and (2) show the full sample, while the remaining columns split the sample into urban, columns (3) and (4), and rural (5) and (6). The data is split in order to allow for the controls to have heterogeneous affects on the urban and rural context. Column (1) shows that the females in the rural setting are significantly less inclined to compete than the males. This is evidenced by the negative estimated coefficient of 0.20, significant at the one percent level, showing that rural females are 20 percentage points less competitive than their male counterparts, who compete 37 percent of the time (as seen from the constant term). The estimated coefficient on Urban is close to zero, and statistically insignificant, showing that urban males are not more competitive than rural males. The interaction term Female Urban is large and statistically significant at the five percent level (p=0.025), showing that females in the urban setting are significantly more competitive than females in the rural setting. Adding the coefficient on Female and this interaction term (Female+Female Urban), we find that females in the urban context are somewhat less competitive than males in the urban context, the estimated coefficient being minus 0.06, but that this difference is not statistically significant (p=0.224). These results in column (1) simply confirm the evidence shown in Figure 1 in the introduction. Column (2) adds controls from the lab, which are likely to affect competition choice, that is, attitude towards risk, knowledge, and confidence. As we have seen from Figures 4 6 these differ between location, gender, or both. Does context matter for the competitiveness of females even when controlling for these dimensions? The answer is: yes. We observe that the estimated coefficients on risk, knowledge and confidence all have the expected sign, and are statistically significant, but that the interaction effect Female Urban is almost unaffected by the inclusion of these controls. Evidently, there is something about being female in the rural context 8 The appendix also splits the data according to gender, rather than location (Table A2) and shows marginal effects of probit regressions rather than OLS (Table A3). 12

13 that makes them less willing to compete, beyond what can be explained by their lower level of knowledge and lower confidence. 9 Table 2. Competition, gender, and context Pooled data Urban Rural (1) (2) (3) (4) (5) (6) Female 0.20*** 0.16*** *** (0.04) 0.14*** Urban 0.03 Female Urban 0.15** (0.07) ** (0.07) Risk 0.04* (0.02) Knowledge 0.02** Confidence 0.02*** 0.07** (0.03) * 0.00 (0.03) 0.04*** 0.02** Constant 0.37*** (0.03) 0.18*** (0.07) 0.40*** (0.04) 0.28*** (0.10) 0.37*** (0.03) 0.10 (0.09) Female + Female Urban R Observations Note: The table reports OLS regressions with Competition (0 1) as outcome variable. Female is an indicator variable that takes the value of 1 if the participant is a female and 0 if a male; Urban is an indicator variable taking the value of 1 if the participant goes to school in the urban location (Kampala), and 0 otherwise (Tororo); Female Urban is an interaction term showing Female multiplied with Urban; Risk shows the number of risky choice made by the participant; Knowledge shows the number of correct answers on the first round of general knowledge questions; Confidence shows how well the participant believes he/she performed compared to the 9 Table A1 in the appendix shows that adding additional controls does not change these conclusions. 13

14 others in the session. Standard errors in parentheses, where *p<0.10, **p<0.05, ***p<0.01. In columns (3) and (4) we run separate regression for the urban context, without and with controls, respectively. We observe that there is no statistical gender difference in competitiveness in this context. Columns (5) and (6) do the same thing in the rural context, showing that females in the rural setting are far less competitive than their male counterparts. The location specific regressions thus confirm the message from the pooled regressions. Our interpretation of these results is that they are based on a difference in the attitudes towards women, even among the women themselves, between the urban and rural context, and that these differences affect their willingness to compete. As we observed in the introduction, women in the rural area appear to be far less empowered than women in the urban area, as evidenced by attitudes to wife beating and control over income. Other studies have shown that the status of women in society shapes the females willingness to compete, and we believe that this is the mechanism at work here as well. An alternative hypothesis could be that the competitive culture in Kampala is more developed than in Tororo, but if this were true we would also expect the Kampalafactor to be significant for males, which is not the case. Yet another possible explanation is that the most competitive females move to the cities where the education and career possibilities are much better. We do, however, have information regarding whom the participants live with. Therefore, if mobility was a factor, we would expect a larger share of females than males in Kampala to live with people other than their parents. We do not find this to be the case. There are more 14

15 participants in Kampala than in Tororo living without their mother and/or father, but there are no significant gender differences along this dimension. 10 Our argument is that the context, and in particular society s attitude towards women, has a direct effect on the females willingness to compete. An alternative hypothesis is that the context shapes competitive preferences through its effect on other variables, such as knowledge, attitude towards risk, and confidence. In order to investigate this issue, we introduce additional interaction effects. In Table 3, column (1) is simply a restatement of column (2) in Table 2, highlighting the estimated coefficients on Female and Female Urban. Table 3: Competition and gender: Interaction effects (1) (2) (3) (4) (5) Female 0.16*** 0.12* *** 0.04 (0.11) (0.11) Interacted with: Urban 0.14** 0.14** 0.15** 0.13** 0.16** (0.07) (0.07) (0.07) (0.07) (0.07) Risk 0.04 (0.04) 0.04 Knowledge (0.02) (0.02) Confidence Observations Note: The table reports OLS regressions with Competition (0 1) as outcome variable. Variables are defined in Table 2. Standard errors in parentheses, where *p<0.10, **p<0.05, ***p<0.01. Columns (2), (3) and (4) add interaction effects with Risk, Knowledge, and Confidence, respectively, and column (5) includes all interaction effects. The interaction term with 10 In the rural area, 53 percent of the females and 56 percent of the males live with their mother and/or father. In the urban area, the corresponding numbers are 36 percent and 35 percent. 15

16 Urban is virtually unaffected by the inclusion of the other interaction effects. Moreover, all of the additional interaction effects are small in size, and none are statistically significant. This suggests that the geographical context, urban versus rural, has a direct effect on females willingness to compete, and that the heterogeneity that we observe in knowledge, risk attitude, or confidence cannot explain the large difference in willingness to compete between urban and rural females. 5. Conclusion Our findings show that females willingness to compete is contingent on context: while females in the rural area compete less than half as much as males, females in the urban context compete just as much as males. Males are equally competitive in both locations. Adding controls, and multiple interaction effects, confirm these results, and suggest that they are driven by the context rather than by differences in knowledge, risk preferences or confidence between the females in the two locations. While our data does not allow us to say exactly how context drives the difference in females willingness to compete, we speculate that cultural factors are likely to be important. In particular, national survey data suggests that women in the urban location are more empowered than their counterparts in the rural location. Given the importance assigned to competitiveness for labor market outcomes, understanding how attitudes towards competition differ between males and females in an urban and rural context of a developing country can be important in the design of anti poverty policies. Our results suggest that promoting females willingness to compete in the rural context may be particularly important. Working for gender equality is important in its own right, but may also have the additional benefit of increasing female s willingness to compete, thereby increasing their economic potential. 16

17 Our results are based on a comprehensive lab experiment with a large number of participants. True, the participants are relatively young and are all in secondary school, and our results may not hold for, say, older people or out of school teenagers. More experiments are needed to say something about the generality of our findings. However, secondary school students are clearly an interesting target group for policy makers and donors. Secondary school students will soon have to make important decisions about work, business and further education, and the school provides a platform for reaching out with messages of empowerment, messages that may shape the competitiveness of the students, and thereby improve their future livelihoods. References Berge, O. L I., Bjorvatn, K., Pires, A. J. G., Tungodden, B., Competitive in the lab, successful in the field? forthcoming Journal of Behavior & Organization Bjorvatn, K., and Tungodden, B., Disabled by stereotype? Experimental evidence from Uganda, forthcoming Journal of Behavior & Organization Booth, A., Nolen, P., Choosing to compete: How different are girls and boys? Journal of Behavior & Organization 81 (2), Croson, R., Gneezy, U Gender differences in preferences. Journal of Economic Literature 47 (2), Douoguih, K. C., Essays in experimental economics with implications for economic development. PhD dissertation, Deparment of Economics. College Park: Faculty of the Graduate School of University of Maryland. Eckel, C.C., Grossman P.J., Males, females and risk aversion: experimental evidence. In Handbook of Experimental Economics Results, red. P. C. og S. V. New York: Elsevier. Gneezy, U., Leonard, K.L., List, J., Gender differences in competition: evidence from a matrilineal and patriarchal society. Econometrica 77 (5),

18 Leibbrandt, A., U. Gneezy, and J. A. List (2013). Rise and fall of competitiveness in individualistic and collectivistic societies, PNAS Early Edition: 1 4. Niederle, M., Vesterlund, L Do females shy away from competition? Do males compete too much? Quarterly Journal of Economics 122, UBOS Uganda National Household Survey 2009/2010, Socio Economic Module [Online]. Kampala: Uganda Bureau of Statistics. Available: [Accessed ]. UBOS Uganda Demographic and Health Survey [Online]. Kampala: Uganda Bureau of Statistics. Available: [Accessed ]. Zhang, J., Narrowing the gender gap in competitive inclination: evidence from the communist experiment in China. Mimeo, Hong Kong University of Science and Technology Zhang, J Can experimental economics explain competitive behaviour outside the lab? Mimeo, Hong Kong University of Science and Technology. 18

19 Appendix Table A1. Competition, gender, and context: additional controls Pooled data Urban Rural (1) (2) (3) (4) (5) (6) Female 0.16*** 0.17*** *** 0.17*** Urban 0.01 Female Urban 0.14** (0.07) 0.01 (0.11) 0.12* (0.06) Risk 0.04** (0.02) 0.04* (0.02) 0.07** (0.03) 0.07** (0.03) 0.00 (0.03) 0.00 (0.03) Knowledge 0.02** 0.02* *** 0.03** Confidence 0.02*** 0.02*** 0.02* 0.02** 0.02** 0.02** Age Disability 0.03 (0.04) 0.05 (0.03) 0.01 (0.06) 0.01 (0.06) 0.07 (0.04) 0.10** (0.04) TV 0.01 (0.04) 0.02 (0.04) Meat (0.02) 0.03 (0.02) 0.03* (0.02) Priming 0.03 (0.03) 0.02 (0.03) (0.04) 0.39 (0.42) Female + Female Urban Session fixed effects No Yes No Yes No Yes R

20 Observations Note: The table reports OLS regressions with Competition (0 1) as outcome variable. Variables are defined in Table 2, and in addition: Age shows the participant s age in number of years. Disabled takes the value of 1 if participant has a disability. TV is an indicator variable taking the value of 1 if the participant has a TV at home. Meat is the number of times the participant eat meat in a week; Priming takes the value 1 if the participant were subject to a disability prime (see Bjorvatn and Tungodden, 2014), and zero if not. Standard errors in parentheses, where *p<0.10, **p<0.05, ***p<

21 Table A2. Competition, by gender Pooled data Female Male Urban (1) (2) (3) (4) (5) (6) ** 0.156*** (0.0335) (0.0445) (0.0501) Risk ** * (0.0207) (0.0271) (0.0310) Knowledge *** *** ( ) (0.0116) (0.0135) Confidence *** *** ( ) ( ) (0.0108) Constant 0.317*** *** *** (0.0167) (0.0569) (0.0219) (0.0713) (0.0248) (0.0943) R Observations Note: The table reports OLS regressions with Competition (0 1) as outcome variable. Variables are defined in Table 2. Standard errors in parentheses, where *p<0.10, **p<0.05, ***p<

22 Table A3. Competition, gender, and context: Probit Pooled data Urban Rural (1) (2) (3) (4) (5) (6) Female 0.23*** 0.19*** *** (0.04) 0.14*** Urban 0.03 Female Urban 0.18** (0.07) ** (0.07) Risk 0.04** (0.02) Knowledge 0.02** Confidence 0.02*** 0.07** (0.03) * 0.00 (0.03) 0.04*** 0.02** Female + Female Urban (0.13) Observations Note: The table reports the marginal effects based on probit regressions with Competition (0 1) as outcome variable. Variables as defined in Table 2. Standard errors in parentheses, where *p<0.10, **p<0.05, ***p<

23 Appendix (not intended for publication) Instructions for workshop 1. Before the session starts [MODERATOR ensures that the participants follow the rules of conduct after entering the room] [When Moderator receives a sign from the Head of the Experiment, he starts reading the introduction] [The session] 2. Introduction Welcome. We appreciate your willingness to participate in this session, which I will lead. In this session you will be asked to make some economic choices, and you will earn money based on your choices and your performance. The results from this session will be used in a research project on entrepreneurship among the youth in Uganda. It is therefore very important that all of you follow certain rules. You are not allowed to talk to any of the other participants during the session. If you have any questions or need any help, please raise your hand and one of us will assist you. We kindly request that both participants and assistants turn off their mobile phones now. If someone does not follow these instructions, we will have to ask him or her to leave the workshop. 23

24 If you need to go to the bathroom during the workshop, please raise your hand. Importantly, do not leave the room without permission. [MODERATOR proceeds when everyone is back from the bathroom] The session will be conducted under anonymity. It will not be possible for the other participants or anyone else, except for the researchers, ever to find out what choices you make, and hence what you earn in the session. You will not be informed about how much money you have earned until the end of the session. The researchers will then prepare an envelope containing the money you have earned, where they will ensure that it is impossible to identify the amount of money inside the envelope simply by looking at it. This envelope will be handed over to you before you go home. It is very important that you remember your desk number and report it in each activity, so that we can pay you correctly. 3. Section Entrepreneurship: Primed or non primed Before we start, we would like to ask you some background questions on entrepreneurship. You will not earn money from answering these questions, but please answer each question as carefully as possible. [MODERATOR waits as sheet titled Entrepreneurship is handed out] You can now turn the sheet. You have 6 minutes to answer the questions. 24

25 [MODERATOR alerts the participants after 3 minutes] [MODERATOR waits 6 minutes] We will now collect the sheets. Make sure that you fill in your desk number. [MODERATOR waits until the sheet Entrepreneurship has been filled in and collected] 3.2 Risk choice Now we move to the next part of the session, where you will be able to earn some money. We will now hand out the sheet that explains the exercise. [MODERATOR waits until the sheet Risk is handed out] You can now turn the sheet, and I ll explain the details. As you can see there are 3 cases on this sheet, all with 2 different alternatives. You have to tick only one box in each case. Only one of the cases will be randomly selected to determine your earnings. Please take a look at the first case. Here you can either choose to receive 2000 Ush guaranteed or you can choose to gamble. If you choose to gamble you can win 4000 Ush, but there is a 25 percent chance of winning and a 75 percent chance of losing, which means you get nothing. 25

26 Is everyone with me? Please raise your hand if you have a question. Now you can tick off your answer in case 1. [MODERATOR waits while everyone make their first choice] Has anyone not finished the first case? [MODERATOR continues when everyone has finished] Please take a look at the second case. Here you can either choose to receive 2000 Ush guaranteed or you can choose to gamble. If you choose to gamble you can win 4000 Ush but there is a 50 percent chance of winning and a 50 percent chance of losing, which means you get nothing. Is everyone with me? Please raise your hand if you have a question. Now you can tick off your answer in case 2. [MODERATOR waits while everyone makes their second choice] Has anyone not finished the second case? [MODERATOR continues when everyone has finished] Please take a look at the third case. Here you can either choose to receive 2000 Ush guaranteed or you can choose to gamble. If you choose to gamble you can win 4000 Ush but there is a 75 percent chance of winning and a 25 percent chance of losing, which means you get nothing. Is everyone with me? Please raise your hand if you have a question. Now you can tick off your answer in case 3. 26

27 [MODERATOR waits while everyone makes their third choice] Has anyone not finished the third case? [MODERATOR continues when everyone has finished] The assistants will now collect your sheets. 3.3 Distribution game 1: Dictator We now move to the next part of the session. My colleagues will hand out the sheet explaining the exercise. [MODERATOR waits until the sheet Distribution 1 has been handed out to all participants, and then continues reading.] You may now turn over the sheets, and I will read the instructions. You are now paired with another participant. You will not know who you are paired with; only the researchers will know this. You and the other participant have together received 2000 Ush in this part of the session. You are now asked to decide how much of the money to give to the other participant. You will receive the rest of the money. You will find examples on your sheets. Please raise your hand if this is not understood. [MODERATOR waits until this is explained to everyone who has raised their hand.] 27

28 You are now to decide how much to give to the other participant. Please fill in an amount between 0 Ush and 2000 Ush. [MODERATOR waits until the sheet Distribution 1 has been collected from everyone.] You have now sent money to another participant. Likewise, another participant has decided how much of the 2000 Ush to give to you. This is not the same person as the one who received the money you sent. 4. Section Objective questions: Round 1 We will now move to the next session. We will ask you some general questions. There are 10 questions, and for each question you can choose between four different answers. Tick off the correct answer. If you tick off more than one alternative, your answer will be considered incorrect. For each correct answer, you are paid 200 Ush. Are you with me? If not, please raise your hand and the assistants will help you. [MODERATOR waits until everyone has been assisted] We will now hand out the questions, but please do not turn over the page before you are told to do so. 28

29 [MODERATOR waits until the sheet Objective Questions Round1 has been handed out. MODERATOR continues when HE gives signal] Remember to fill in your desk number, so that we can pay you correctly. You have 10 minutes to answer the questions. You can now turn over the sheet. [MODERATOR alerts the participants after 5 minutes] [MODERATOR waits until HE gives a signal 1 minute left] [MODERATOR waits until HE signals that the sheets can be collected] This part of the session is now over. Please lay down your pens. My assistants will collect the sheets. 4.2 School and leisure: Primed or non primed We would now like to ask you some background questions related to school and leisure activities. You will not earn money from answering these questions, but please answer each question as carefully as possible. [MODERATOR waits until the sheet School and Leisure has been handed out] You can now turn the sheet, you have 6 minutes to answer the questions. 29

30 [MODERATOR alerts the participants after 3 minutes] [MODERATOR waits 6 minutes] We will now collect the sheets. [MODERATOR waits until the sheet School and Leisure has been collected] 4.3 Beliefs and competition choice We will now move to the next session. The assistants will hand out the sheets. [MODERATOR waits until the sheet Beliefs and Competition choice has been handed out] You may now turn the sheets. On the top half of the page, titled Beliefs, we first ask you how many you think you got right out of the 10 Objective Questions you answered before the questions about school and leisure. Second, we ask how many on average you think the others got right. We give a reward of 1000 Ush for each of the guesses that turn out to be correct. Look at the bottom half of the sheet, titled Competition choice. We will soon give you a second set of Objective Questions. This time, however, we will give you a choice between two different kinds of payment. Your first option is to receive 200 Ush for each correct answer. Your second option is to enter into a competition. Then you will be paid 500 Ush per correct answer if you are at least as good as the average in the first Objective Question test. If you provide fewer correct answers than this average, you will receive nothing. 30

31 Is this understood? If not, please raise your hand. [MODERATOR waits until everyone has been assisted and continues reading] Please tick off the one alternative you prefer. [MODERATOR continues when everyone has finished] The assistants will now collect the sheets. 4.4 Distribution game 2: Spectator choice In the next part of the session, you will make a decision that could have real consequences for two other participants in the session. I will explain the decisions in detail, but first the assistants will hand out the sheet explaining the exercise. [MODERATOR waits until the sheet Distribution 2 has been handed out to all participants, and then continues reading] You are now to make a decision involving two other participants. You will not know who these two participants are. The two participants have earned an additional reward from answering the Objective Questions. The size of this reward depends on their performance in answering those questions. One of the participants performed well and ended up among the top 50%. He or she has earned the high reward of 1500 Ush. The other participant performed not so well, and ended up among the bottom 50%. He or she earned the low reward of 500 Ush. Together, these two participants have therefore earned a total reward of 2000 Ush. Your job is to decide how to distribute this total reward between the two 31

32 participants. You can choose between splitting the additional rewards equally or giving each of them the reward he or she earned. You should now decide how to split the money between the two participants, by ticking off one of the boxes on your sheet. [MODERATOR waits until the sheet Distribution 2 has been filled in and collected] 4.5 Objective questions: Round 2 The assistants will now hand out the second set of Objective Questions. You will work according to your chosen rate, either a fixed rate or a competition rate. [MODERATOR waits until the sheets Objective Questions Round 2 has been handed out to all participants] You have 10 minutes to answer the questions. You can now turn over the sheets. [MODERATOR alerts the participants after 5 minutes] [MODERATOR waits until HE gives a signal 1 minute left] [MODERATOR waits until HE signals that the sheets can be collected] This part of the session is now over. Please lay down your pens. My assistants will collect the sheets. 5. Section Trust game 32

33 In the next part of the session, you are paired with another participant. You will not know who you are paired with. You and the other participant will have different roles. One of you will first make a decision and then the other participant will respond to this. We will now hand out a sheet to those of you who are to make the first choice; we ask the rest of you to wait patiently. [MODERATOR waits until the sheet Send Money has been handed out to all senders (ODD NUMBERS), and then continues reading] You may now turn over the sheet. You are given 2000 Ush, whereas the participant you are paired with is given nothing. You can now decide to send some of the 2000 Ush to the other participant. The other participant receives three times the amount you send. You will find examples on your sheet. The other participant will later determine how much of what he or she has received to return to you. Please raise your hand if this is not understood. [MODERATOR waits until this is explained to everyone who has raised their hand] You are now to decide how much to return to the other participant. In the box at the bottom of the sheet, please fill in an amount between 0 and 2000 Ush. 33

34 [MODERATOR waits until the sheet Send money has been filled in and collected] We will now hand out a sheet to those of you who are to respond. We ask the rest of you to wait patiently. [MODERATOR waits until the sheet Return money (EVEN NUMBERS) has been handed out to all returners, and then continues reading] You may now turn over the sheet. The other person was given 2000 Ush, whereas you were given nothing. You have received three times what the other participant sent. You will find examples on your sheet. In the first box on your sheet, you can see how much you have received. You are to decide how much of this you want to return to the other participant. Please raise your hand if this is not understood. [MODERATOR waits until this is explained everyone who has raised their hand] Please fill in an amount between 0 Ush and the amount you have received. [MODERATOR waits until the sheet Return money has been filled in and collected] 6. Section Use Language: Primed or non primed 34

35 We will now ask you to do a language exercise. You will not earn money from this exercise, but please work on it as carefully as possible. Importantly, fill in you desk number. [MODERATOR waits as the sheet Use Language has been handed out ] You can now turn the sheet, and I ll explain the exercise. Each of you has in front of you a list of words. For each set of words, please make a sentence with four words. Use only four out of the five words. You will find an example on your sheet. You have 6 minutes to work on the exercise, please start now. [MODERATOR alerts the participants after 3 minutes] [MODERATOR waits 6 minutes] We will now collect the sheets. [MODERATOR waits until the sheet Use Language has been filled in and collected] 6.2 Multiple choice question: Round 3 We will now move to the next part of the session, where we hand out the third and final round of Objective Questions. This time everyone competes. 35

36 You will be paid 500 Ush per correct answer if you are at least as good as the average in the first Objective Question Test. However, if you provide fewer correct answers than this average, you will receive nothing. Is this understood? The assistants will now hand out the final set of Objective Questions. You have 10 minutes to answer the questions. You can now turn over the sheets. [MODERATOR alerts the participants after 5 minutes] [MODERATOR waits until HE gives a signal 1 minute left] [MODERATOR waits until HE signals that the sheets can be collected] This part of the session is now over. Please lay down your pens. My assistants will collect the sheets. 6.3 Time Choice In this part of the session you will not earn money. Imagine if you could choose between receiving an amount of money early or a higher amount later. My colleagues will hand out the sheet that explains the exercise. [MODERATOR waits until the sheet Time choice has been handed out to all participants.] 36

37 You may turn over your sheet, and I ll explain this in detail. As you can see, the early payment is 1 week from now, and the later payment 5 weeks from now. The early payment is 1000 Ush, while the later payment is 3000 Ush. Is this clear? Please tick off the one alternative you prefer. [MODERATOR waits until the sheet Time choice has been filled in and collected.] 6.4 Beliefs 2 We would now like to ask you some questions on how well you think you and other participants, both people with disabilities and non disabled, performed in the last objective question test. We will now hand out the sheet. [MODERATOR waits until the sheet Beliefs 2 has been handed out] You may now turn the sheets. [MODERATOR waits until the sheet Beliefs 2 has been filled in and collected] 7. Section 5: Final questions 7.1 Career Choice Now we would like you to answer some questions about career choice. 37

38 My assistants will now hand out the questions. [MODERATOR waits until the sheet Career Choice has been handed out] First we will look at question 1. Assume that you could choose between the following job opportunities; private sector employee, government employee, own business and farmer, and that the income and work hours were exactly the same in all of them. You should rank them so that 1=most preferred, 2=preferred second, 3= preferred third and 4=least preferred. Has everyone answered question 1? [MODERATOR waits until everyone has finished] Now look at question 2. What would you do if you had 1 million Ush? You can choose A. use them to buy something nice for myself or my family. B. Use them to start a business. C. Use them to pay my education. If you choose D. Others, please specify in the box below. Only tick one of the boxes. Question 3 is a bit long. Here we want you to picture yourself 10 years from now. Look at the career choices and answer each and everyone from a scale from 1 to 5, where 1 is very unlikely and 5 is very likely. Take your time to tick off one choice per career option. Has everyone finished answering question 3? [MODERATOR waits until everyone has finished] Now look at the final question, question 4. Where do you see yourself living in 5 years from now? Like question 3, answer the likelihood of living in a big city, small town, rural area and outside Uganda on a scale from 1 to 5 where 1 is very unlikely and 5 is very likely. Tick only one choice per location. 38

39 Has everyone finished question 4? [MODERATOR waits until everyone has finished] My colleagues will now collect your sheets. 7.2 Background information Now we would like you to answer some background questions. The assistants will now hand out the final questions. [MODERATOR waits until the sheet Background has been handed out] You may now turn the sheets. [MODERATOR waits until the sheet Background has been filled in and collected] We would like to thank you all for participating in this session. Your input will be most valuable for our research project on entrepreneurship among the youth in Uganda. May we ask you not to discuss this session with others before the end of the next two weeks, since we will arrange further sessions with other students in Tororo in the coming days. Please leave the pen on your desk when you leave the room. Again, thank you for your participation in this workshop. 39

40 [THANK YOU!] After lunch you will receive your payment. Now, please follow my colleague. [Break... ] [Participants enter the room and the MODERATOR starts reading ] Please find your desks. We will now start handing out the envelopes starting with the row in front. Please bring your payment form and form a line. [Whenever one row is finished the MODERATOR asks the next row to line up] 40