3 Comparison with current STCs

For this recruitment drive, we implemented an assessment test which consisted of 6 questions (2 questions by level: beginner, intermediate, and advanced) for Stata, R, and Python. The assessment test was only provided when the applicant mentioned that she has some level of said statistical program. While these 6 questions cannot provide a complete evaluation of the applicants’ skill, it gives an idea of their skill level, and we also tested these questions with the current STCs in order to validate our instrument. The goal behind this assessment test is to provide more insights in the recruitment process, so TTL can pick better their STCs.

3.1 Overall self-assessment

As stated above, we first asked the applicants what they think their software skill level is before we provided them with the built-in assessment questions. Figure 3.1 and 3.2 shows the percentage of women and men self-assessted skill level.

Stata self-assessment by gender

Figure 3.1: Stata self-assessment by gender

R self-assessment by gender

Figure 3.2: R self-assessment by gender

On the other hand, figures 3.3 and 3.4 presents the Stata and R results, respectively, from the assessment test by gender.

Stata assessment results by gender

Figure 3.3: Stata assessment results by gender

R assessment results by gender

Figure 3.4: R assessment results by gender

3.2 Are the current RAs and FCs better than the applicants?

Short answers, yes. We tested the assessment questions with the current RAs and FCs at DIME. Overall, the current RAs are more likely to answer the questions correctly and had a better score on average for both Stata and R.

3.3 Raw differences between current RAs and applicants

3.4 Stata Test

3.5 R Test