Relevant education and rigorous assessment
Kåre Moberg, Senior Researcher for the Danish Foundation for Entrepreneurship-Young Enterprise Denmark
Is it possible to assess the impact of entrepreneurship education? Some would say that this is fairly simple whereas others believe it to be impossible. Many researchers take a position somewhere in between.
Without throwing any rocks, I would say that I am fairly sceptical about the way in which many assessment studies are performed today, but I do believe that an increased focus on the activity is crucial. There are numerous problems to face when we attempt to perform assessment studies, and it is next to impossible to address them all since it is human activity, intertwined in complex social settings, and affected by a myriad of internal and external factors, that we as educational programme evaluators try to study. It is difficult to argue that simple cause-effect relations are possible in such a setting.
So… what can we actually study, assess and evaluate? This is an ongoing debate within the research community which is often referred to as the relevance vs. rigour debate. On one side, there are researchers who argue that the gold standard for rigorous assessment studies is randomized controlled trials (RCT). On the other side, we have the researchers who claim that the meticulous rigour of the RTC methodology renders it useless for studying anything relevant. In RCTs the participants who get the educational “treatment” are randomly selected and matched with a control group (an elaborate presentation of this method can be found here and here). Critics of the RCT methodology say that it is more or less impossible to control for all factors influencing educational outcomes, and therefore often argue that a different sampling method should be used. Instead of getting a small amount of information from a large number of respondents, their claim is that a study that focuses on a smaller number of respondents but dedicate more time with each respondent would generate more relevant insights. The participants selected for such a study would typically be the participants that were affected the most by the educational initiative – a so-called extreme sampling method (interesting perspectives on this can be found here and here)
The importance of randomization
While I agree that the latter approach may generate relevant and valuable insights about how educational initiatives influence participants I do not see how these types of studies can demonstrate the effectiveness of different educational initiatives. As long as the educational “treatment” has not been randomly distributed there are often problems with “self-selection bias” which is turned into a problem with “survival bias” by the extreme sampling method applied. It therefore becomes impossible to compare the effectiveness of different educational initiatives.
Even if a course or a programme is mandatory rather than elective it is still very difficult to assess its impact and influence. This has to do with the fact that it is more or less impossible to control for all factors influencing educational outcomes (I know, I am nagging about this) regardless of whether the data has been collected with short questionnaires or extensive and lengthy qualitative interviews. The only way around this problem is to use randomization. If the sample size is sufficient and the chance of getting the educational treatment is equal, then all other factors influencing the educational outcome will also be distributed at random. It will thus be equally likely that factors influencing the educational outcome occur in both of the groups (experiment group and matched control group). Events influencing for example entrepreneurial intentions, attitudes and self-efficacy, such as friends or family members starting companies or experience bankruptcy are equally likely to occur to participants receiving and participants not receiving the educational “treatment”.
Addressing the challenges
There are a lot of assessment tools available today that both evaluators and practitioners can use (an overview of examples can be found here). By combining these tools it is possible to answer a broad scope of different research questions. So, it is no longer an issue of how we measure the outcomes, rather how we structure the data collection.
Naturally, the method applied should be determined by the questions the study seeks to answer, and usually a mixed method approach would generate the most interesting insights. However, when it comes to assessing the effectiveness and efficiency of different educational initiatives, it is, in my view, necessary to follow the foundational principals of the RCT method, however difficult this might be. As evaluators, we need to be creative when designing our programme evaluations. A way to use randomization while making sure that all participants are equally rewarded is to use an “in-phase” randomization method, that is, randomly selecting who gets the educational treatment first and who gets it next. If the randomization is performed at the institutional level it would also be possible to follow the participants in a longitudinal way as it will be a new set of participants that function as treatment and control groups.
Should we bother?
So, is it really that important that we perform rigorous and relevant assessment studies of educational initiatives? Well, according to a survey that the UN started in 2015, the almost 10 million participating respondents held education to be the most important topic out of 16 possible areas, spanning from access to clean water, jobs and healthcare, to democracy and climate issues. I would thus say that we have an obligation to perform rigorous evaluations that generate relevant insights. It will not be easy, but in order to further our understanding about how educational initiatives are best designed and delivered it is important that we rise to the challenge.Category : entrepreneurship education, impact research Posted : 1 June 2016 08:34 UTC