Entrepreneurship training programmes are an important component of demand-side job creation strategies in developing countries (Fox and Kaul, 2017). Assessments of such programmes are
constrained by variations in the programme content, as entrepreneurship training is often combined with grants, life-skills training, internships and mentorship. The targets of these programmes also vary and include vulnerable groups, subsistence entrepreneurs as well as firms which have greater potential for growth. The indicators of success should be adapted to suit the objectives and target group of the programme. Given the varied nature of entrepreneurship training programmes, it is unsurprising that a range of indicators are used to assess them. The indicators can be grouped into three broad categories: indicators of business practices, indicators of business performance and psychological indicators. Income and profits are the most commonly used indicator (Cho & Honorati, 2014). Randomised control trials which compare treatment and control groups are the gold standard method for assessing entrepreneurship training programmes. However, the quality of these studies can be improved by having larger sample sizes, baseline assessments before the intervention and three to four follow-up assessments to assess the long-term success of the programme.
Entrepreneurship training programmes vary in terms of content, length and target groups. Most programmes combine entrepreneurship training with cash grants, microfinance, life-skills training, vocational training, internships or mentorship. The programmes target necessity entrepreneurs, firms with better prospects for growth and vulnerable groups such as women or marginalised youth (Valerio, Parton, & Robb, 2014). Given these differences between the programmes, it is difficult to compare the success of the programmes. There are several meta-analyses and
systematic reviews which compare the impact assessments of several entrepreneurship training programmes. These studies reveal that a range of indicators has been used to assess the success of the programmes (Glaub & Frese, 2011; McKenzie & Woodruff, 2013; Patel, 2014). The indicators can be grouped into three broad categories: indicators of business practices, indicators of business performance and psychological indicators. Business performance indicators, especially income and profits, are the most widely used type of indicators (Cho & Honorati, 2014).
Entrepreneurship training programmes for women may require specialised content such as life skills training or mentorship that can assist women to overcome social barriers to their participation in the labour market (Patel, 2014). Programmes which target women tend to include psychological indicators which measure improvements in self-confidence, decisionmaking, empowerment or agency.
Randomised control trials which are an experimental research design that compares treatment groups which are exposed to intervention to a control group which is not exposed are the best
method for assessing the impact of entrepreneurship training programmes (Cho & Honorati, 2014; Glaub & Frese, 2011). The comparison of the treatment and control groups eliminates the
effect of extraneous variables which the study cannot control for, such as macroeconomic changes. The randomised control trial is more effective if the sample size is larger because the
statistical analysis will have greater precision and the study has more power to detect the small effects of changes which may occur after the intervention. It is necessary for this kind of study to
have a baseline assessment of the key indicators before the intervention and three to four follow-up assessments which can evaluate short-term and long-term effects. Studies which utilise this
approach can determine if the short-term spike in self-employment rates, income or profits which typically occur after training are sustainable over a longer time period (McKenzie & Woodruff,
2013). Programmes and impact assessments which have a longer time frame usually have to contend with participant drop-out or attrition. Randomised allocation of the sample to treatment or
control groups is a critical strength of the randomised control trial, however complex, multi-phase programmes which expect participants to graduate from one phase to the next must be designed
to cope with attrition (D. J. McKenzie & Puerto, 2017). An impact assessment should be designed to ensure that the findings are not influenced by the Hawthorn effect, which occurs when respondents give positive responses after the training because this is deemed to be socially desirable. The most sophisticated programme assessments test for displacement effects which may occur when improvements observed after training come at the expense of groups which did not receive training (McKenzie & Woodruff, 2013).