However, the weakness of this negative correlation raises the notion that the results have no correlation at all. To investigate this possibility I will first try to find how just how weak this negative correlation is. As you can see from graph 2 the different result values for each temperature vary quite a lot from the average. This undermines the accuracy of the results, as the greater the range of the results from the average the less likely the results and therefore the average is accurate.

The range of the different results for each temperatures and the standard deviation from the mean can indicate how accurate the results are. This is because, as I have explained earlier the range and the deviation of the results from the mean determines how accurate the results are. Graphs 3 and 4 show how the standard deviation and the range of the results change over the temperature range used in the experiment.

In graph 3 the range was calculated by subtracting the lowest value from the highest, and in graph 4 the standard deviation was calculated using the formula; ? ((? (x -? x)/n) (where x = the values and n = the number of values). Graph 3 and 4 show that in general the accuracy of the result increases with an increase in the temperature, and that at low temperatures the results are not very accurate at all.

I can conclude this as, the bigger the standard deviation or range the less accurate the results are. In both graphs 3 and 4 the general pattern shows high range and high standard deviation at low temperature (high inaccuracy) and low range and low standard deviation at high temperatures (low inaccuracy). This therefore contradicts my earlier theory that the inaccuracy would increase with an increase in the temperature, however it must be taken into account that the results are very inaccurate already.

Because of this inaccuracy the results may have had a certain random element in them, evidence for this is in the fact that the results are very inaccurate (they have a large range and a large standard deviation), and that the inaccuracy varies a lot from its trend line (see graphs 3 and 4). So in conclusion, I think that the results of this experiment were not accurate at all and not accurate enough to draw any sound conclusions that are backed with scientific theory/knowledge.

In fact I think there was a certain random element in the results that contributed to the inaccuracy of the results. Because of the inaccuracy of the results, I feel that it is not a good idea to use these results as a basis for any conclusions, especially as these conclusions cannot be backed up by any scientific knowledge. Therefore with only the results I have at this moment, I have to conclude that the resistance is directly proportional to the temperature. I can conclude this because I have decided to use scientific knowledge as a basis for the answer to this correlation.

I have done this because the results I have are too inaccurate to use for making any judgements on this correlation. Evaluation: So far I do not think this investigation has gone very well because of the very inaccurate results, however with the large amount of scientific knowledge I have been able to come to a conclusion to this investigation that has a lot of backing. My prediction was of corse correct. The reason to this inaccuracy in the results is, in my opinion due to the large amount of contacts and wiring used in this method.

Some components react to temperature by changing the resistance in different ways, most components react in a directly proportional linear way, but some components like a filament lamp or diode react in a directly proportional but logarithmic way and/or only at certain voltages. It is quite likely that in the multi-meters there were some of these components that react in a logarithmic way and at certain voltages. This means that the current flowing through the circuit would be affected out of proportion by these components, therefore affecting the resistance.

Although the affect of these components would have been neutralised if one multi-meter were used, it might not be neutralised if two multi-meters were used as they were in this method. The large amount of wire used in this experiment has meant the current could have been easily affected by other variables that I did not take into account, like a change in the room temperature, or if I accidentally spilt some water onto the wires. Also like the different depths used in the water bath, or a magnetic flux from the high voltage used in the kettle.

Due to the large amount of outside variables that can affect the large amounts of wire used, I have to conclude that the large amounts of wire used in this method was the main cause of the inaccuracy of its results. I cannot estimate how accurate my results are because I have no basis on what the true values are, only that the correlation is directly proportional. If I were to repeat this investigation, I would change the method so that less wire was used and only one multi-meter was used, I would also change the method so as to stop any of the variables mentioned in this evaluation affecting the results.

If possible I would try and use a more accurate multi-meter in my experiment. A more accurate multi-meter will give me more accurate results; the reduced amount of wire to be used will reduce the effect of outside variables and make the results more reliable. I dont think I need to do any extra research, as I have more than enough scientific backing for any conclusions I will have to make if I repeat this investigation.

The range of the different results for each temperatures and the standard deviation from the mean can indicate how accurate the results are. This is because, as I have explained earlier the range and the deviation of the results from the mean determines how accurate the results are. Graphs 3 and 4 show how the standard deviation and the range of the results change over the temperature range used in the experiment.

In graph 3 the range was calculated by subtracting the lowest value from the highest, and in graph 4 the standard deviation was calculated using the formula; ? ((? (x -? x)/n) (where x = the values and n = the number of values). Graph 3 and 4 show that in general the accuracy of the result increases with an increase in the temperature, and that at low temperatures the results are not very accurate at all.

I can conclude this as, the bigger the standard deviation or range the less accurate the results are. In both graphs 3 and 4 the general pattern shows high range and high standard deviation at low temperature (high inaccuracy) and low range and low standard deviation at high temperatures (low inaccuracy). This therefore contradicts my earlier theory that the inaccuracy would increase with an increase in the temperature, however it must be taken into account that the results are very inaccurate already.

Because of this inaccuracy the results may have had a certain random element in them, evidence for this is in the fact that the results are very inaccurate (they have a large range and a large standard deviation), and that the inaccuracy varies a lot from its trend line (see graphs 3 and 4). So in conclusion, I think that the results of this experiment were not accurate at all and not accurate enough to draw any sound conclusions that are backed with scientific theory/knowledge.

In fact I think there was a certain random element in the results that contributed to the inaccuracy of the results. Because of the inaccuracy of the results, I feel that it is not a good idea to use these results as a basis for any conclusions, especially as these conclusions cannot be backed up by any scientific knowledge. Therefore with only the results I have at this moment, I have to conclude that the resistance is directly proportional to the temperature. I can conclude this because I have decided to use scientific knowledge as a basis for the answer to this correlation.

I have done this because the results I have are too inaccurate to use for making any judgements on this correlation. Evaluation: So far I do not think this investigation has gone very well because of the very inaccurate results, however with the large amount of scientific knowledge I have been able to come to a conclusion to this investigation that has a lot of backing. My prediction was of corse correct. The reason to this inaccuracy in the results is, in my opinion due to the large amount of contacts and wiring used in this method.

Some components react to temperature by changing the resistance in different ways, most components react in a directly proportional linear way, but some components like a filament lamp or diode react in a directly proportional but logarithmic way and/or only at certain voltages. It is quite likely that in the multi-meters there were some of these components that react in a logarithmic way and at certain voltages. This means that the current flowing through the circuit would be affected out of proportion by these components, therefore affecting the resistance.

Although the affect of these components would have been neutralised if one multi-meter were used, it might not be neutralised if two multi-meters were used as they were in this method. The large amount of wire used in this experiment has meant the current could have been easily affected by other variables that I did not take into account, like a change in the room temperature, or if I accidentally spilt some water onto the wires. Also like the different depths used in the water bath, or a magnetic flux from the high voltage used in the kettle.

Due to the large amount of outside variables that can affect the large amounts of wire used, I have to conclude that the large amounts of wire used in this method was the main cause of the inaccuracy of its results. I cannot estimate how accurate my results are because I have no basis on what the true values are, only that the correlation is directly proportional. If I were to repeat this investigation, I would change the method so that less wire was used and only one multi-meter was used, I would also change the method so as to stop any of the variables mentioned in this evaluation affecting the results.

If possible I would try and use a more accurate multi-meter in my experiment. A more accurate multi-meter will give me more accurate results; the reduced amount of wire to be used will reduce the effect of outside variables and make the results more reliable. I dont think I need to do any extra research, as I have more than enough scientific backing for any conclusions I will have to make if I repeat this investigation.