The differentiation has not been made by the originator of the thread - are we talking about standard or modified proctor? A 95% modified proctor may be more difficult to get than a 98% standard proctor, depending on the soils. Looking at QLA, a 98% is certainly going to have *much* less qualitative failure than a 95%.
It depends on the penalty for failure. What are you supporting?
LCruiser has a point in that the "type" of standard (modified (heavy) or standard (light)) was not stated. Rather than how the specified value was chosen - i.e., 95% Mod MDD or 98% Mod MDD, I think the question is what is the real behaviour difference between a material compacted at either of the two efforts. And as important what kind of material with which we are dealing. The differences may be more remarkable in clayey soils where the effects of porewater pressures during compaction can have an effect than in gravelly sands or well graded crushed stone. In the latter, the material hardness also plays a part in that higher compactive effort will break down the aggregate - or knock off the sharp angled edges thereby reducing the interlock.
I have not seen anything definitive to say that at 98% MDD for a given material (say a well graded gravelly sand (SW)) will settle only 11 mm under a given footing load whereby if the compaction was to 95%, it would settle 14, 15 or 16 mm.
Sadly (or happily depending on the point of view) chosen levels of compaction by different engineers is based on experience, judgment, what the company has always done successfully and the like. There is scant few real case histories. As I indicated earlier, you cannot, say, take the behaviour at a specified level of 95% and believe that it will reflect the actual behaviour - insofar as one layer might have been 96%, a second layer 95%, the third 98%, the fourth 94% (and does it really mean anything so long as the entire fill layer is 95% or more??). These are the philosophical questions, in my mind.
One more aspect to consider: Just what is the accuracy of a field density test? My opinion is that specifying 98 percent comaction probably is what it takes to get the outliers to fall above 95 percent compaction. I'm fully confident that the precision of any field density test is not to the single unit.
On the matter of Modified v. Standard - I've always taken it as a 5 percent difference. 95 modified is about 100 standard. Actually the modified proctor test was developed to keep folks from worrying when they calculated over 100 percent compaction (can't of course have over 100 percent, eh?)
fattdad hit my point a bit better than I originally did. I believe the reason for specifying higher percentage compaction is often (and primarily) to counter what is considered to be poor quality control or even the inability to check compaction on every square foot of the project. If you are only checking compaction every 1,000 cubic yards, you have a lot of fill that is not being checked. Some of those areas will not meet your specified densities. The material properties may not match the samples tested in the lab either. By specifiying that additional 3% density as an acceptance criteria, the engineer has greater assurance that the entire fill has been placed to a "minimum" of 95%. Basically, you are going to pay the contractor to beat on it a bit more so you can rest assured that the entire mass is closer to your desired criteria.
I am cognizant of the means of specifying, why it is done (rightly and wrongly), the variations, etc. but, if I have the same material at 95% MDD (Std or Modified) and compare it to 98% MDD (corresponding Std or Modified) - what is the difference in phi? in deformation modulus? That's the original question I believe. The relationship of Std vs Modified (light vs heavy for the Brits) varies according to the material. I have had it in the order of 3% - 100% std MDD = 97% mod MDD for an Ontario crushed stone subbase course - but can be more so up to 5% doesn't seem unreasonable.
I think what you are really thinking of is How to correlate 95%/98% proctor compaction to N value, or results of plate load test or CPT. It would answer your question about how little more 98% can carry than 95% and how much can 95% carry.
There may well be such correlations somewhere which I will be very interested to know too!
I practice in the Piedmont province where a large portion of the near-surface soils are fine-grained and contain a good portion of clay and silt (CL, MH and ML). As part of an internal experiment and since we were somewhat slow in our work load, I used a pressuremeter to test controlled fill sites to determine the soil bearing capacity and settlement potential. The fill sites were graded under our observation to a minimum 95% Standard Protor MDD at +/- 3% of optimum moisture content. We would create a pilot hole with a specialized shelby tube and run the pressuremeter test within the newly created pilot hole. The shelby tube would then be transported back to the lab to determine % compaction. We tested over 20 sites. With the exception of a few soft layers (no grading contractor is perfect), the pressuremeter results indicated the presence of soils capable of supporting 3,000 psf. The soils had compactions ranging between 95 to 100% Standard Proctor MDD. What this means is that, in my opinion and for the soils in my geographic area, I do not see the need to spec higher than 95% Standard Proctor WITH MOISTURE CONTROL LIMITS for conventional foundation support for single to low-story construction.
We have typically seen geotechs recommend 92 to 95 % when compacting fill for footings to bear on up to 6 or 8 feet in depth. For large amounts of fill say 20 feet or so, we have seen geotechs calling for 98%. I assume that was spec'd to control the total amount of settlement.
