We failed to reject the null hypothesis. Our study found that there was no significant difference in the weather forecast accuracy between Accuweather and Weather.com. However, both achieved just over 50% success rate. The forecast accuracy indicates the two weather apps may not be the most useful for UBC students and residents. Perhaps the coastal weather is more unpredictable than other climates, or the meteorological methods used by these providers are not very good. Further studies are needed to see if other weather website services can provide more accurate 12 hour forecasts.

There are several potential limitations to the study we conducted. The first one was time restriction. Due to insufficient access to past weather forecast in UBC area as well as inadequate time to conduct this study, we were not able to gather and evaluate weather information in the four seasons at UBC. This could negatively affect our results as the accuracy of weather forecast could vary with the changing seasons. A study done by Lupo and Market (2002) shows that seasonal variation in weather forecast accuracy does exist. For example, the National Weather Service (NWS) gives better predictions about weather in autumn and winter than in spring and summer. Another limiting factor was sample size as we were not able to compare the weather forecast with real-time weather hourly due to our course load. However, as we measured five times a day with a fixed length of time separation, the opposing effect brought about by this should be minimized.

Moreover, the sampling method may carry the possibility of experimental error. We observed the weather and compared it to the prediction made by the weather forecast service. Since the weather can be quite variable within an hour, there exists the possibility that the time of observation did not represent the entire hour. In addition, it was difficult to tell whether it is partly cloudy or partly sunny, making it hard to determine true or false for a given observed weather condition. Last but not least, different group members may have made different judgments over weather conditions. To minimize these limitations we standardized the sampling methods and criteria.