For a certain electronics retailer during each of eight days

Note that the correlation depicted in the table is not positive correlation. Positive correlation means that if one variable increases, the other increases too and if one decreases, the other decreases too. But what we see is a predominantly negative correlation. When television bar increases/decreases, the warranty bar decreases/increases in most data points. The exception is Day 4 when both decrease together. If I were to remove day 4, the table will depict a negative correlation.
The scatter plot has all the points and the trendline showing positive correlation only because of the first point on the left (circled in Red). This point is the data of Day 4. Remove this point and your trendline changes to the one shown in green. It is negatively correlated. (shown in the attachment) 
Hence the ANSWER here will be 'Day 4' and 'No longer positively correlated'.
You can read about Correlation and Trendlines in my Data Insights Module. It will be freely available tomorrow (Sunday) under the Super Sundays program. Details of Super Sundays available here.                                     
­

 KarishmaB Abhivas­ 
For a certain electronics retailer during each of eight days, both of the graphs show the number of televisions sold and the number of extended warranties sold (whether for televisions or for any other products sold at the store). The scatterplot also includes a trendline that shows the correlation between the daily numbers of televisions and warranties sold.
Select from the drop-down menus the options that complete the statement so that it is accurate based on the information provided.
Among the days represented in the graphs, without the data for Day Select146 ,  the daily numbers of televisions and extended warranties sold would Selectno longer be positively correlatedno longer be negatively correlatedstill be negatively, though less strongly, correlatedstill be positively, though more strongly, correlated .


 Note that the correlation depicted in the table is not positive correlation. Positive correlation means that if one variable increases, the other increases too and if one decreases, the other decreases too. But what we see is a predominantly negative correlation. When television bar increases/decreases, the warranty bar decreases/increases in most data points. The exception is Day 4 when both decrease together. If I were to remove day 4, the table will depict a negative correlation.
The scatter plot has all the points and the trendline showing positive correlation only because of the first point on the left (circled in Red). This point is the data of Day 4. Remove this point and your trendline changes to the one shown in green. It is negatively correlated. (shown in the attachment) 
Hence the ANSWER here will be 'Day 4' and 'No longer positively correlated'.
You can read about Correlation and Trendlines in my Data Insights Module. It will be freely available tomorrow (Sunday) under the Super Sundays program. Details of Super Sundays available here.                                     
­Hi @KarishmaB @chetan2u
I understood why 'Day 4' and 'No longer positively correlated' is the answer but i am confused about why Day 1 and still be positively, though more strongly, correlated will not be the answer.

Below is my understanding:  
Because of Day-1 data, graph is going downwards, but if we remove this Day-1 data, we are removing one negative slope and graph will have a positive slope for MOST of days. Please help!!



­  

Can someone explain why Day 6 is not an outlier. What would happen if we removed day 6 from the data?
chetan2u

Note that the correlation depicted in the table is not positive correlation. Positive correlation means that if one variable increases, the other increases too and if one decreases, the other decreases too. But what we see is a predominantly negative correlation. When television bar increases/decreases, the warranty bar decreases/increases in most data points. The exception is Day 4 when both decrease together. If I were to remove day 4, the table will depict a negative correlation.
The scatter plot has all the points and the trendline showing positive correlation only because of the first point on the left (circled in Red). This point is the data of Day 4. Remove this point and your trendline changes to the one shown in green. It is negatively correlated. (shown in the attachment) 
Hence the ANSWER here will be 'Day 4' and 'No longer positively correlated'.
You can read about Correlation and Trendlines in my Data Insights Module. It will be freely available tomorrow (Sunday) under the Super Sundays program. Details of Super Sundays available here.                                     
­

­KarishmaB I totally agree with the first part, as according to the approximate coordinates of the trend - line [(17,15) the lower (30,20) the upper] and the 3 days listed d1(15,27) d4(12,15) d6(25,20) we can see that d1 has the smaller deviation (x=2 y=7) and it's excluded, d4 has x=5 y=0 and the d6 is inside (for x) and at the upper limit (for y) so the d4 is the outlier. The scatter plot also points out that 4th day is the only one away from the rest and as a result is the one that affects the most. But for the second part, after also drawing a scatter plot it seems that again the correlation is negative just not so much. Because from day to day when one coordinate for instance decreases (let's say d1 to d2) the other either is the same (d1 to d2) or increases (d3 to d5). Can you help me please on this?­

­
By drawing the trendline with a positive slope, they are implying a positive correlation (though a weak positive correlation). We cannot say that the correlation was negative to begin with because of the trendline in the scatter plot and the mix of data in the bar graph. 

After you remove day 4, you do get a clear negative correlation in the bar graph and in the scatter plot and that is why we choose "no longer positively correlated." When one increases the other decreases and vice versa. Look at the data from one day to the next day in the bar graph.