Management decided to introduce an incentive plan to try to increase the productivity in a small manufacturing plant. The morale of the workers was measured before and after the introduction of the new plan. A scale of 1 to 10 was used, where 1 is a very low morale and 10 a very high morale. Can we say that the new incentive plan did improve the morale of the workers? ( Use α=0.025 and do a sign test).

Before 8 7 6 9 7 10 8 6
After 6 5 8 6 9 8 10 7

1 answer

To determine whether the new incentive plan improved the morale of the workers, we can use the sign test as a non-parametric statistical test. The sign test will help us assess whether there is a significant difference in the morale ratings before and after the introduction of the incentive plan.

Step 1: Set Up Hypotheses

  • Null Hypothesis (\(H_0\)): There is no difference in morale before and after the incentive plan. (The median difference is 0.)
  • Alternative Hypothesis (\(H_1\): There is a difference in morale before and after the incentive plan. (The median difference is not 0.)

Step 2: Calculate Differences

We need to calculate the differences between the "Before" (B) and "After" (A) scores for each worker: \[ \text{Difference} = A - B \]

| Before | After | Difference (A - B) | |--------|-------|---------------------| | 8 | 6 | -2 | | 7 | 5 | -2 | | 6 | 8 | 2 | | 9 | 6 | -3 | | 7 | 9 | 2 | | 10 | 8 | -2 | | 8 | 10 | 2 | | 6 | 7 | 1 |

Step 3: Count Signs of the Differences

Next, we count the number of positives and negatives:

  • Negative Steps: -2, -2, -3, -2 -> Total = 4
  • Positive Steps: 2, 2, 2, 1 -> Total = 4
  • Zero Differences: None

Step 4: Conduct the Sign Test

The sign test uses the smaller of the number of positive and negative differences. Here, both are 4. Thus, we have \( N = 8 \) (total non-zero differences).

We find the critical value for a two-tailed test at \(\alpha = 0.025\). The critical value for a binomial distribution with \(N = 8\) at a significance level of \(\alpha\) can be found in binomial tables or calculated using binomial probabilities.

The critical region for the sign test at \(\alpha = 0.025\) is at least 1 or the values that fall below the 2.5% threshold on either tail (with 8 observations):

  • This means we will reject \(H_0\) if the number of positive signs \( < 1 \) or the number of negative signs \( < 1 \).

Using binomial probabilities, we first calculate the probabilities of the number of successes (here, positive differences).

\[ P(X = k) = \binom{N}{k} p^k (1-p)^{N-k} \]

Where:

  • \(N = 8\) is the total number of differences
  • \(k\) is the number of positive differences (successes)
  • \(p = 0.5\)

For critical values, we will evaluate if the count of signs is less than or equal 1:

Calculating: \[ P(X = 0) = \binom{8}{0} (0.5)^0 (0.5)^8 = \frac{1}{256} \approx 0.0039 \] \[ P(X = 1) = \binom{8}{1} (0.5)^1 (0.5)^7 = 8 \times \frac{1}{256} = \frac{8}{256} = \frac{1}{32} \approx 0.03125 \]

Adding these gives: \[ P(X \leq 1) = P(X=0) + P(X=1) \approx 0.0039 + 0.03125 = 0.03515 \]

Since this is greater than 0.025, we fail to reject the null hypothesis.

Conclusion

At a significance level of \(\alpha = 0.025\), there is not enough evidence to conclude that the new incentive plan significantly improved morale in the manufacturing plant. The morale of the workers did not statistically change.