Initially, our group decided to establish a relationship between peak timings (at MRT stations) and length of waiting times (for trains) as well as duration of waiting times as we wanted to do something different. Four of the group members were assigned to note down the duration of timings of trains travelling towards Joo Koon, Pasir Ris, Jurong East and Marina Bay on two days, 14 June and 16 June 2010. It was tedious and since the recordings were done manually using different stopwatches/timers, there were bound to be inaccuracies in the recorded timings. The results were computed into the SPSS system. However, the Cramer's V test could not be applied since there was only one continuous variable and when the Pearson's Correlation Test was applied, a strong linear correlation between the peak timings and duration of waiting times could not be attained.
For example, the tables below show the data obtained for the respective train timings on 14 June.
(PH: Peak Hour, OPH: Off Peak Hour)
(JK: Joo Koon, PR: Pasir Ris, MB: Marina Bay, JE: Jurong East)
(1.00: length of time greater than 2 mins, 0.00: length of time less than or equal to 2 mins)
Down in the dumps.
It was quite demoralizing for the whole group to be informed by the lecturer herself that we had deviated from the requirements. Many thoughts and emotions filtered through our minds and it was clear we had been affected by this. The lecturer advised us to stick with our data and the direction to which our project was then headed and we contemplated it. However, it was only right that we discussed over our next course of action in length and revised our goals.
The turning point.
Following a unanimous vote, our group decided to work on the correlation between a person's BMI and his height whereby height and BMI/weight would be the scale variables. We had our hearts set to rectify our past mistakes. It was like having a second wind!
Here we go again!
On the day of data collection, it was ensured that the measuring tape was placed against the wall as vertically as possible so as to maximise the accuracy of the height values as possible. An electronic weighing machine was used for measuring weight. In general, it was sometimes tough to get students to participate as some of them may be sensitive about having their particulars, weight and height taken. The data was collected within NYP campus from students within the SHS block, specifically outside LTH1. The students' respective ages and genders were noted down before the measurements (height and weight) were taken. As initially there were more females students who participated in the survey, attempts to get an almost equal number of male students (by getting more male students from north canteen) were carried out to ensure an even distribution of results. Furthermore, more than 30 samples were taken to minimise skewness of data due to outlier value (weight of obese individual).
We missed whattttt???
It was unfortunate that during the process of taking height and weight measurements, we failed to consider that some of the students' weights were taken after their lunch. As a result, there might have been inaccuracies in the weight values as their weights might had been higher than usual.
The thought process.
The data was computed and the Pearson Correlation Test was applied to see correlation between the BMI and height values. However, there was not a strong positive linear relationship between the two variables. Nonetheless, a complete study could be conducted this time round. It was definitely a welcome respite amidst our already hectic schedule.
It's over?? YAY!!!
It was an enriching and interesting experience for us to be involved in this project. We experienced disappointment at the start and near delirium with the end in sight. It certainly taught us that patience and perseverance will reap dividends in the end. Furthermore, we have learnt more about how the statistical tests could be applied in practical situations/scenarios.
