Question 1

A researcher in a Human Resources Unit presented a recent study, which showed a statistical significance between length of staff lunch breaks and low productivity; how can she explain to her manager that this does not mean that the length of staff lunch breaks should be reduced?a.

Accepted Answer

A significant result only suggests that there is a relationship between the length of staff lunch breaks and low productivity, but it does not mean that this relationship is important enough to warrant reducing lunch breaks for all staff. Other factors, such as work hours and workload, may also contribute to low productivity, and reducing lunch breaks may not necessarily address the root cause of the issue. Therefore, it is important to take a holistic approach to addressing low productivity rather than relying solely on one statistical result.

Question 2

You have just joined the sales modelling team for a start-up software company. Your boss has decided that from now on the team will adopt a Bayesian approach. However, not all staff understand what this is; your boss asks you to present a training session. How would you explain a Bayesian approach in your session introduction?

Accepted Answer

A Bayesian approach involves updating the likelihood of a statistical model as more data is collected. This allows for a more accurate analysis and interpretation of the data. It is different from other approaches, such as frequentist or classical statistics, which do not update the likelihood of the model as more data is collected.

Question 3

A stockmarket trader conducted a Bayesian analysis of variations in skirt length and stock market growth. He calculated a Bayes factor of 1. Should he use skirt length as a predictor of stock market growth?

Accepted Answer

A Bayes factor of 1 indicates that the evidence does not favor one hypothesis over the other, meaning that skirt length variations do not provide useful information for predicting stock market growth.

Question 4

A trainee data analyst for a large social media company, which has falling site usage, has just completed a study into factors that affect site users' satisfaction levels. However, he finds only one statistically significant factor, which he includes in his report but he deliberately, omits the other six non-significant findings. What is the term for what the data analyst has done?

Accepted Answer

P-hacking refers to the practice of manipulating data or experimenting until nonsignificant results become significant. However, the scenario described is more accurately an example of omitting non-significant findings rather than manipulating data to achieve significance. The correct term for this practice is not provided among the options, which might involve terms like "cherry-picking" or "selective reporting." Nonetheless, given the options, none directly describe the act of omitting non-significant findings as described, but p-hacking is often associated with practices that could lead to biased or misleading statistical conclusions, which might be why it was chosen. However, it's important to clarify that the precise action described in the question does not match the definition of p-hacking.

Question 5

Your business studies lecturer has devoted the past ten weeks to teaching you the Bayesian approach and is now asking that you offer a critique of it. What key criticism could you raise?

Accepted Answer

The key criticism of the Bayesian approach is that the reliance on a prior probability is overly subjective and can be open to a researcher's degrees of freedom. This means that the choice of prior probability can greatly influence the final outcome and may be biased by the researcher's own views and beliefs. This subjectivity can be problematic, particularly if the prior probability is not well-informed or is based on limited data.

Question 6

Your CEO has followed your advice and now wants you to measure effect sizes. You report a Pearson's r of 0.50 for the impact of Unblock Me Now drain cleaner on reducing drain blockage time. Your CEO wants to know if this is bad, as she remembers that a p-value of 0.30 is not good. What do you tell her?

Accepted Answer

P-value and effect size are not the same things. An effect size of r=0.50 is generally considered large and suggests that the Unblock Me Now drain cleaner has a substantial impact on reducing drain blockage time. This means that the product is promising, and the CEO should consider moving ahead with the launch of Unblock Me Now.

Question 7

A recruitment analyst wanted to examine the likelihood that advertising on social media is more effective than in print media for recruiting the best candidates. She conducted one study where the probability of making a Type I error was 0.05 and a Type II error was 0.2. Does her research have empirical probability?

Accepted Answer

Empirical probability refers to the probability obtained through repeated identical experiments rather than just a single study. Therefore, option B is the correct answer as it acknowledges the need for repeated identical experiments while also considering the specific probabilities of Type I and Type II errors.

Question 8

Your CEO has just read a book on criticisms of the NHST and worries that all company data analysis is now flawed and will lead to huge financial losses. How might you reassure her?

Accepted Answer

It is true that NHST has its flaws, but completely dismissing it as a method for data analysis may not be the best solution. Incorporating an examination of effect sizes can help us understand the magnitude of the effect and the practical significance of the findings. This can help us interpret the results more accurately and make informed decisions based on the data.

Question 9

You are the CEO of a small financial forecasting company. You have decided to adopt a Bayesian approach to data analysis and modelling. When you announce this new policy, your staff are unhappy and unconvinced, as they are used to a NHST approach. You stress that the Bayesian approach has several key advantages, including which of the following.

Accepted Answer

The Bayesian approach allows for the evaluation of the likelihood of the null hypothesis being true, which is not possible in the NHST approach. This can provide more nuanced and informative conclusions about the data being analyzed. None of the other options listed are advantages of the Bayesian approach.

Question 10

You lead a product-testing unit for a large pharmaceutical company. Your team has conducted forty trials of a new antibiotic but you are not sure if the results are conclusive enough to urge the company to start producing the new drug. A new data analyst has joined your team suggesting that meta-analysis might be a good idea, do you agree?

Accepted Answer

Meta-analysis would be a good idea in this case because the forty trials were identical and tested the same research question, allowing for the calculation of an average effect size for the new drug. This can provide more conclusive results than looking at the individual trials alone.

Question 11

A researcher was assessing customer satisfaction with MakeMebeautiful, a new beauty product. He had a sample size of 75 and a p-value of 0.10. Does the researcher recommend that the company stop promoting this product?

Accepted Answer

A small sample size can greatly affect the accuracy and reliability of p-values. Therefore, the researcher cannot confidently recommend that the company stop promoting the product based on the small sample size and the marginally significant p-value.

Question 12

Which of the following is not a factor in data analysts' over-use of p-values and NHST in business and marketing research?

Accepted Answer

Statistical software does not encourage over-use of p-values or NHST. In fact, it is important for data analysts to understand the assumptions and limitations of NHST and to use it appropriately in conjunction with other forms of analysis. Time constraints, pressure to get significant results, and company bonus structures are all factors that may contribute to over-reliance on NHST and p-values in business and marketing research.

Question 13

In our previous example, the human resources manager had already calculated the probability of no women being hired based on sector wide data. In the Bayesian approach, what sort of probability is this?

Accepted Answer

The probability calculated based on sector wide data is a prior probability, which is used as a starting point for Bayesian analysis.

Question 14

A human resources manager in the IT sector was concerned about unconscious bias in recruitment panels. There were two posts and seven candidates, four men and three women. Theoretically, all the candidates have an equal probability of being hired as they all match the selection criteria. However, the manager has data that suggests that it is more likely men will be hired based on data from across the IT sector and within her own company. However, the manager has implemented many equality initiatives within her company and therefore wants to determine the probability that still no women will be hired. What formula could she use to determine this probability and assess the impact of unconscious bias in her company's recruitment?

Accepted Answer

Bayes' theorem is a statistical formula that can be used to calculate the conditional probability of an event, given prior knowledge of related events. In this case, the HR manager can use Bayes' theorem to calculate the probability of no women being hired, given the historical data and potential unconscious bias. This formula can help with assessing the impact of bias in recruitment and determining the effectiveness of any interventions or initiatives implemented to promote equality. NHST, Pearson's r, and Cronbach's Alpha are not relevant in this scenario.

Question 15

A researcher working in a Human Resources department was interested in gender and sales figures so he conducted a t-test. The mean for males was 66.25 and the mean for females was 78.24, with both groups having a standard deviation of 7. What is the effect size using Cohen's d?

Accepted Answer

Cohen's d formula: 
d = (M1 - M2) / S
Where M1 and M2 are the means of Group 1 and Group 2, respectively, and S is the pooled standard deviation. 
Using this formula, we can calculate:
d = (78.24 - 66.25) / 7
d = 1.714
The effect size using Cohen's d is 1.712 (rounded to three decimal places), which indicates a large effect size according to Cohen's guidelines. Therefore, the best answer is A.

Question 16

An HR manager was interested in employee use of company on-site gyms across twenty sites. Different researchers collected and analysed data across each of the sites but the resultant twenty reports showed differing p-values, some sites found a statistical significance between opening hours of on-site gyms and employee usage and others did not. Which of the following would it useful for her to review?

Accepted Answer

The answer of An HR manager was interested in employee...

Quiz 3: The Phoenix of Statistics