In: Statistics and Probability
In her paper titled “Choking under pressure of top performers: Evidence from biathlon competitions”, Florian Lindner looks at the effect of pressure on performance in Biathlon World Cup competitions. Her main regression involves regressing the variable ‘missed fifth shot’ which is a dummy variable equal to 1 if the shot was unsuccessful and 0 if the shot was successful (missing the decisive fifth shot is defined as choking under pressure) on several variables that include:
‘lane’-This variable serves as a proxy for relative performance, i.e., the rank at the beginning of the final shooting bout (the fastest competitors and those with the highest probability to finish in the top positions occupy shooting lanes with lower numbers). But this variable is normalized from 0 to 1 so that 0 corresponds to the highest lane number and 1 corresponds to the lowest lane number.
‘shtime’-This refers to the shooting time. The author argues that longer shooting times are probably related to less automatic execution in a precision task, which might be linked to overthinking, speculating about possible outcomes and losing one’s focus. This potentially increases the probability of choking under pressure, i.e., missing the final shot.
The sign of the coefficients on the independent variables give the direction of the effect on the probability that the dependent variable is 1. For instance, a coefficient of -1 on ‘ability’ means that an increase in ability leads to a decrease in the probability of missing the fifth shot.
The results are being presented below (standard errors are in the parentheses below the coefficients). Consider column 4.
1. Calculate the t statistic for the coefficient on the variable ‘lane’.
2. Is there evidence to support the following hypothesis: Athletes who experience higher levels of pressure due their higher position in the ranking have a higher probability of experiencing deterioration in performance. Explain.
3. Is there evidence in support of the hypothesis that longer shooting times lead to an increase in the probability of missing the fifth shot? Explain
4. Does gender play a role in determining the probability of missed fifth shot?
5. Are these results evident of the phenomenon of ‘choking under pressure’?
1.
t statistic for 'lane' = estimated coefficient / Standard error
= 1.849 / 0.240 = 7.704166667
2.
Higher positon in ranking implies low lane number and high value of lane number. As, the variable 'lane' is significant (shown by *** signs) in model 4 and the sign of the coefficient is positive, there is high probability of missing fifth shot with increase in lane. Hence, there is evidence to support that the "athletes who experience higher levels of pressure due to their higher position in the ranking have a higher probability of experiencing deterioration in performance".
3.
As, the variable 'shtime' is significant (shown by ** signs) in model 4 and the sign of the coefficient is positive, there is high probability of missing fifth shot with longer shooting times. Hence, there is evidence to support that the "longer shooting times lead to an increase in the probability of missing the fifth shot".
4.
As, the variable 'male' is not significant (no * signs) in model 4, the gender does not play a role in determining the probability of missed fifth shot.
5. As, the p-value of the model (Prob > ) is 0.000 which is less than the assumed significance level of 0.05, the model is statistically significant. Hence, there is significant evidence of 'choking under pressure'.