Hypothesis test?

J

johnjacob

New member
Joined
Nov 14, 2013
Messages
5
Hello, I am trying to use a hypothesis test with this quarterly information of units produced within a standard turnaround time but i can't figure out which one to use. Any help appreciated. The table is similar to this for each quarter except the number of units produced each quarter is different. Any idea how I should approach this problem or which hypothesis to use?

Receiving Processes - Quarter 1
Units Produced During Quarter: 9,500
ProcessPercentage Processed within Turnaround Time StandardsVariance from Turnaround Time Standards
Supplier arrive8,4551,045
Supervisor verify materials8,740760
Validate materials to be received8,645855
Team unload truck8,0751,425
Take raw material to plant8,2651,235
Log given to clerk at the end of the day8,550950
Clerk enter info into system8,740760
 
johnjacob said:
Hello, I am trying to use a hypothesis test with this quarterly information of units produced within a standard turnaround time but i can't figure out which one to use. Any help appreciated. The table is similar to this for each quarter except the number of units produced each quarter is different. Any idea how I should approach this problem or which hypothesis to use?

Receiving Processes - Quarter 1
Units Produced During Quarter: 9,500
ProcessPercentage Processed within Turnaround Time StandardsVariance from Turnaround Time Standards
Supplier arrive8,4551,045
Supervisor verify materials8,740760
Validate materials to be received8,645855
Team unload truck8,0751,425
Take raw material to plant8,2651,235
Log given to clerk at the end of the day8,550950
Clerk enter info into system8,740760
Click to expand...
I don't think the column labeled "Variance" is the customary use of that word. Since the sum of the two columns is equal to the Units Produced, it looks like a table oh how many were within the standard time and how many were not. A possible hypothesis could be that these are all samples from a binary distribution with p=0.894 and n=9500. That would not be a GOOD hypothesis, because the standard deviation of the binary distribution would be much smaller than the variations.

What is different for the other quarterly reports? Is it possible that there are "always" 89% that exceed standard time?

What have you tried?
 
DrPhil said:
I don't think the column labeled "Variance" is the customary use of that word. Since the sum of the two columns is equal to the Units Produced, it looks like a table oh how many were within the standard time and how many were not. A possible hypothesis could be that these are all samples from a binary distribution with p=0.894 and n=9500. That would not be a GOOD hypothesis, because the standard deviation of the binary distribution would be much smaller than the variations.

What is different for the other quarterly reports? Is it possible that there are "always" 89% that exceed standard time?

What have you tried?
Click to expand...

Yes it is just the difference as you mentioned. Not the real variance. The other three quarters only difference from this table is the units produced during the quarter. Therefore, the units processed also differ from quarter to quarter.

I was thinking that for each process i could find the mean and variance. Then create a 5-step hypothesis testing. I would like to create a hyphothesis say that if we improve the first process "Supplier arrive" then all other process would improve. but I am not sure how to execute that. I am not sure about a binary distribution. I am trying to learn t- Z- and ANOVA.
i appreciate your help.
 
johnjacob said:
Yes it is just the difference as you mentioned. Not the real variance. The other three quarters only difference from this table is the units produced during the quarter. Therefore, the units processed also differ from quarter to quarter.

I was thinking that for each process i could find the mean and variance. Then create a 5-step hypothesis testing. I would like to create a hyphothesis say that if we improve the first process "Supplier arrive" then all other process would improve. but I am not sure how to execute that. I am not sure about a binary distribution. I am trying to learn t- Z- and ANOVA.
i appreciate your help.
Click to expand...
You have data for 7 processes and for 4 quarters.. Somebody has specified how much time "should" be used for each (independent) process. If those "standard times" are appropriate (the null hypothesis?), then the numbers (or fractions) not meeting the standard should be random and not correlated. Is there one process that always has more failures? Is there enough evidence to reject a null hypothesis that the standard times are accurate?
 
DrPhil said:
You have data for 7 processes and for 4 quarters.. Somebody has specified how much time "should" be used for each (independent) process. If those "standard times" are appropriate (the null hypothesis?), then the numbers (or fractions) not meeting the standard should be random and not correlated. Is there one process that always has more failures? Is there enough evidence to reject a null hypothesis that the standard times are accurate?
Click to expand...

I still don't correlate what I can do with this info. The standard times given are as follows and the receiving process percentage for Q1...

Receiving Processes
In Process and Out of Process Standards
Process
 In Process Standards
 Out of Process Standards
Supplier arrive
 <= 14 days
 > 14 days
Supervisor verify materials
 <= 72 hours
 > 72 hours
Validate materials to be received
 <= 72 hours
 > 72 hours
Team unload truck
 <= 48 hours
 > 48 hours
Take raw material to plant
 <= 48 hours
 > 48 hours
Log given to clerk at the end of the day
 < End of Business Day
 > End of Business Day
Clerk enter info into system
 <= 24 hours
 > 24 hours


Receiving Processes – Quarter 1
Process
 Percentage Processed within Turnaround Time Standards
 Variance from Turnaround Time Standards
[FONT=Tahoma, Calibri, Verdana, Geneva, sans-serif]Supplier arrive[/FONT]
89%
 (11%)
[FONT=Tahoma, Calibri, Verdana, Geneva, sans-serif]Supervisor verify materials[/FONT]
92%
 (8%)
[FONT=Tahoma, Calibri, Verdana, Geneva, sans-serif]Validate materials to be received[/FONT]
91%
 (9%)
[FONT=Tahoma, Calibri, Verdana, Geneva, sans-serif]Team unload truck[/FONT]
85%
 (15%)
[FONT=Tahoma, Calibri, Verdana, Geneva, sans-serif]Take raw material to plant[/FONT]
87%
 (13%)
[FONT=Tahoma, Calibri, Verdana, Geneva, sans-serif]Log given to clerk at the end of the day[/FONT]
90%
 (10%)
[FONT=Tahoma, Calibri, Verdana, Geneva, sans-serif]Clerk enter info into system[/FONT]
92%
 (8%)
 
johnjacob said:
I still don't correlate what I can do with this info. The standard times given are as follows and the receiving process percentage for Q1...

Receiving Processes
In Process and Out of Process Standards
Process
 In Process Standards
 Out of Process Standards
Supplier arrive
 <= 14 days
 > 14 days
Supervisor verify materials
 <= 72 hours
 > 72 hours
Validate materials to be received
 <= 72 hours
 > 72 hours
Team unload truck
 <= 48 hours
 > 48 hours
Take raw material to plant
 <= 48 hours
 > 48 hours
Log given to clerk at the end of the day
 < End of Business Day
 > End of Business Day
Clerk enter info into system
 <= 24 hours
 > 24 hours


Receiving Processes – Quarter 1
Process
 Percentage Processed within Turnaround Time Standards
 Variance from Turnaround Time Standards
Supplier arrive
89%
 (11%)
Supervisor verify materials
92%
 (8%)
Validate materials to be received
91%
 (9%)
Team unload truck
85%
 (15%)
Take raw material to plant
87%
 (13%)
Log given to clerk at the end of the day
90%
 (10%)
Clerk enter info into system
92%
 (8%)
Click to expand...
In the 1st quarter, the furthest out-of-standard was "Team unload truck." Is that also true in other quarters? How about using Order Statistics to identify where either procedures need to be improved, or the standard changed?
 
DrPhil said:
In the 1st quarter, the furthest out-of-standard was "Team unload truck." Is that also true in other quarters? How about using Order Statistics to identify where either procedures need to be improved, or the standard changed?
Click to expand...

Yes that is true for other quarters. I just want to create a simple hypothesis that verifies that if I improve the delivery time of supplies arriving would likely trickle down to improvements to the rest of the process.
 
Anyone can give me a hint which type of hypothesis i could use in this case to do a 5-step test?
 
You must log in or register to reply here.
Top